Mortality Predictors in Stroke Patients Requiring Mechanical Ventilation: A Multicenter Prospective Observational Study

Abstract

Background: Patients with acute severe stroke requiring mechanical ventilation represent a significant clinical challenge. Identification of mortality predictors is necessary to improve outcomes. Methods: Fourteen hospitals located around Russia participated in this prospective multicenter observational clinical study. Patients admitted to ICU between November 1, 2017, and November 1, 2019 with confirmed cerebral stroke, aged 18 to 90 years, and requiring mechanical ventilation were included. The impact of various clinical factors on mortality during the 28-day period after stroke was assessed. Results: A total of 1289 patients were included in the registry, and 1144 met the study criteria. The 28-day mortality rate for stroke patients on mechanical ventilation was 64.3%. The most common indications for mechanical ventilation were impaired consciousness (75.7%) and hypoxemia (60.9%). In the cohort of strokes with NIHSS severity greater than 20 points, hypoxemia at the start of ventilation (OR 1.85 [1.21; 2.81], P = 0.004) and the use of hyperventilation mode (OR 1.46 [1.02; 2.06], P = 0.0336) were associated with increased mortality. Pressure-controlled mode as the primary ventilation method (OR 0.36 [0.21; 0.60], P < 0.001) and ICP monitoring (OR 0.23 [0.12; 0.44], P < 0.001) were associated with decreased mortality. Infectious complications were associated with longer mechanical ventilation and ICU stay (P < 0.001). The relationship between probable mortality and the severity of neurological deficit on the NIHSS scale at the start of mechanical ventilation is non-linear. A critical threshold was reached at 16 points NIHSS, where a trend of increasing probable mortality emerged. Conclusion: The identified predictors of mortality in stroke patients requiring mechanical ventilation are essential for decision-making in this cohort. They include hypoxemia, hyperventilation (used to control intracranial hypertension), volume-controlled (VC) versus pressure-controlled (PC) initial ventilation, and the use of clinical methods for monitoring ICP alone versus invasive monitoring.